Ab Initio Study of Nuclear Magnetic Shieldings and Ultraviolet Spectra

ACS2GO © 2019. ← → → ←. loading. To add this web app to the home screen open the browser option menu and tap on Add to homescreen...
0 downloads 0 Views 861KB Size
Download PDF
J. Am. Chem. SOC.1995,117, 11523-11528

11523

Ab Initio Study of Nuclear Magnetic Shieldings and Ultraviolet Spectra for Hypothiocyanite and Its Isomers. The Molecular Structure of Hypothiocyanite Dage Sundholmi Contribution from the Institut f i r Physikalische Chemie, Lehrstuhl f i r Theoretische Chemie, Universitat Karlsruhe, 0-76128 Karlsruhe, Germany Received February 9, 1995@

Abstract: The I3C, 15N, I7O, and 33Schemical shieldings for three optimized structures of hypothiocyanite have been calculated at the SCF, MP2, and CCSD levels using the gauge-including atomic-orbital approach. The first few excitation energies for the corresponding hypothiocyanous acids have been obtained using the equation-ofmotion CCSD and the CI singles methods. Comparison of the calculated 15Nchemical shifts and of the excitation spectra with experimental data suggests that hypothiocyanite has a bent OSCN- structure with C, symmetry, even though the linear SNCO- isomer is 86 kJ/mol lower in energy. The linear ONCS- structure is 146 kTlmol above SNCO-. No other low-lying structures were found. OSCN- and HOSCN are known.8914-16The I5N chemical shift for HOSCN/OSCN- has recently been measured in a phosphate The peroxidase-catalyzedoxidation of SCN- by H202 results buffer solution in the presence of lactoper~xidase.'~ in the formation of a chemical species with antimicrobial Different linear isomers of the hypothiocyanite anionI7 and Several different species such as cyanide (CN-),6 of the corresponding hypothiocyanous acid1*have been studied thiocyanogen ((SCN)Z),~cyanosulfurous acid (H02SCN) or using Hartree-Fock (SCF) and second-order Mdler-Plesset cyanosulfuric acid (H03SCN),8 and hypothiocyanite (OSCN-) (MP2) calculations. In the ab initio study the lowest energy or hypothiocyanous acid (HOSCN)2-4 have been suggested to was found for the linear SNCO- isomer. At the MP2 level, be responsible for the antimicrobial action. The peroxidasethe linear ONCS- structure was found to be 145 kJ/mol above catalyzed oxidation reaction of SCN- by H202 has been SNCO-, and no stable structure was found for the linear OSCNextensively studied experimentally and several chemical methstructure. The MP2 calculations provided two plausible strucods to characterizing the product have been r e p ~ r t e d . ~ *The ~ * ~ - ' ~ tures for hypothiocyanite, but it was not possible to judge from major product of the SCN- oxidation has been identified to be the calculations which isomer is generated in the peroxidasethe hypothiocyanite anion (OSCN-) or the hypothiocyanous acid catalyzed oxidation reaction of SCN-. In the ab initio study (HOSCN).2-4$13%14 The molecule is unstable at higher concenno hypothiocyanite isomers with nonlinear structures were trations and its chemical properties are not well-known. The con~idered.'~ same oxidation product (HOSCN/OSCN-) is also obtained by Since the I5N chemical shift has been measured for OSCN-, the oxidation of (SCN)2 with water.2 However, the oxidation in the presence of lactoperoxidase in a phosphate buffer of KSCN in aqueous solutions by electrolysis or by bromine ~olution,'~ the molecular structure of OSCN- may be determined yields another species which is much more stable than hyby comparing experimental chemical shieldings with the chemipothiocyanite and hypothiocyanous acid (HOSCN/OSCN-).2 cal shieldings calculated for a few plausible molecular structures. The ultraviolet (W)spectra for HOSCN/OSCN- in aqueous Comparison of the calculated excitation energies for the acid solutions has been measured, and the extinction coefficients for form of the isomers with experimental data supports the determination of the molecular structure of the oxidation + On leave from the Department of Chemistry, P.O. Box 55 (A.I. Virtasen product. aukio l), FIN-00014 University of Helsinki, Finland. In this work, two linear structures (SNCO- and ONCS-) and @Abstractpublished in Advance ACS Abstracts, November 1, 1995. (1) Biorck. L.: Rosen. C. G.: Marshall. V.: Reiter. B. Aoul. Microbiol. one bent structure (OSCN- with CI symmetry) of the hypothio1975: 3b, 199. cyanite anion, and the corresponding structures of the hypothio(2) Aune. T. M.: Thomas. E. L. Eur. J . Biochem. 1977. 80. 209. cyanous acid (HSNCO (CJ, HONCS (CJ, and HOSCN (CI)) (3) Hoogendoom, H.; Piessens, J. P.; Scholtes, W.; Stoddard,L. A. Caries have been studied by means of large scale ab initio calculations. Res. 1977, 11, 77. (4) Thomas, L. E. Biochemistry 1981, 20, 3273. The molecular structures have been optimized at the MP2 level ( 5 ) Pruitt, K. M.; Reiter, B. In The Lactoperoxidase System: Chemistry using three different basis sets. The I3C, 15N, I7O, and 33S and Biological Sign$cance; Pruitt, K. M., Tenovuo, J. Eds.; Marcel chemical shieldings have been calculated at the SCF, MP2, and Dekker: New York, 1985; p 143.

1. Introduction

"

(6) Chung, J.; Wood, J. I. Arch. Biochem. Biophys. 1970, 141, 73. (7) Aune, T. M.; Thomas, E. L. Biochemistry 1978, 17, 1005. (8) Hogg, D. M.; Jago, G. R. Biochem. J . 1970, 117, 779. (9) Wilson, I. R.; Harris, G. M. J . Am. Chem. SOC.1960, 82, 4515. (10) Thomas, E. L.; Bates, K. P.; Jefferson, M. M. J . Dent. Res. 1980, 59, 1466. (11) Pruitt, K. M.; Tenovuo, J. Biochim. Biophys. Acta 1982, 704, 204. (12) Pruitt, K. M.; Tenovuo, J.; Andrews, R. W.; McKane, T. Biochemistry 1982, 21, 562. (13) Magnusson, R. P.; Taurog, A.; Doms, M. J . B i d . Chem. 1984,259, 13783.

(14) Modi, S . ; Deodhar, S . S . ; Behere, D. V.; Mitra, S. Biochemistry 1991, 30, 118.

(15) Pruitt, K. M.; Tenovuo, J.; Mansson-Rahemtulla, B.; Hanington, P.; Baldone, D. C. Biochim. Biophys. Acta 1986, 870, 385. (16) Tenovuo, J.; Pruitt, K. M.; Mansson-Rahemtulla, B.; Hanington, P.; Baldone, D. C. Biochim. Biophys. Acta 1986, 870, 377. (17)Pyykko, P.; Runeberg, N. J . Chem. Soc., Chem. Commun. 1991, 547. (18) Runeberg, N. M.Sc. Thesis, Department of Chemistry, University of Helsinki, 1992.

0002-7863/95/1517-11523$09.00/0 0 1995 American Chemical Society

Sundholm

11524 J. Am. Chem. SOC., Vol. 117, No. 46, I995 coupled-cluster (CCSD) level^'^-^^ using the gauge-including atomic-orbital method (GIAO). Transition energies to the few lowest excited states have been calculated at the coupled-cluster level using the equation-of-motion coupled-cluster method (EOM-CCSD).24

Table 1. Optimized MP2 Geometries (Bond Distances, R , in pm and Angles, a, in deg) for the ABCD Isomers of the Hypothiocyanite Anion and the Hypothiocyanous Acids (HABCD) Using Different Basis Sets molecule

basis set

R(AB)

R(BC)

R(CD) a(ABC) a(BCD)

OSCN-"

DZP TZ2P TZVPP DZP TZ2P TZVPP DZP TZ2P TZVPP DZP TZ2P TZVPF TZVPPf TZVPPg

157.68 156.14 155.99 124.57 125.19 124.83 171.72 170.51 169.62 167.92 166.52 165.84 139.10 169.74

173.70 172.70 171.42 119.01 118.00 118.12 118.65 117.96 118.08 170.63 170.14 169.52 122.38 121.47

118.87 118.03 118.34 166.89 166.66 165.95 121.78 121.61 121.67 117.94 117.07 117.33 156.26 117.39

2. Computational Methods In this study, the fully optimized Gaussian basis sets of references 25 and 26 have been used. The Karlsruhe basis sets which have been optimized at the self-consistent-field (SCF) level were augmented by polarization functions. The employed basis sets are the double-l; basis set plus one polarization function (DZP), the triple-l; basis set plus two polarization functions (TZ2P), and the triple5 valence basis set plus two d-functions and one f-function for C, N, 0, and S, and two p-function and one d-function for H (TZVPP). The exponents of the polarization functions of the DZP basis sets are 0.80 for H, 0.80 for C, 1.00 for N, 1.20 for 0, and 0.55 for S. The exponents of the polarization functions of the TZ2P basis sets were obtained by dividing and multiplying the exponent of the polarization function of the DZP basis set by &. The polarization functions of the TZ2P basis sets have the following exponents: 0.46 and 1.39 for H, 0.46 and 1.39 for C, 0.58 and 1.73 for N, 0.69 and 2.08 for 0, and 0.32 and 0.95 for S. The polarization functions of the TZVPP basis set are the polarization functions of Dunning's correlation consistent basis sets (cc-pVTZ).27,28Due to program restrictions in the nuclear magnetic resonance (NMR) chemical shift program, six Cartesian dfunctions have been used in the calculation of the chemical shieldings. The MP2 DZP and M E TZ2P geometry optimizations and the CCSD calculations were for consistency also performed using six d-functions. In the MP2 geometry optimizations using the TZVPP basis sets only the true spherical d- and f-components were included in the basis set. Correlation effects have been considered by using the secondorder M$ller-Plesset perturbation theory (MP2) method, the coupled-cluster method with singles and doubles excitations (CCSD),29 and the CCSD method with perturbative triple excitations (CCSD(T)).30,31The excitation energies have been obtained using the equation-of-motion CCSD method (EOMCCSD)24and the configurationinteraction singles (CIS) method (or the Tamm-Dancoff approximati~n).~~ The I3C, 15N, "0, and 33S chemical shieldings have been calculated at the SCF, MP2,'9-21 and CCSD22%23 levels using the gauge-including atomic-orbital method (GIAO) and the MP2 equilibrium geometries. The DZP and TZ2P chemical shielding calculations, the coupled-cluster (CCSD and CCSD(T)) calculations, the equation-of-motion coupled-cluster (EOM-CCSD), and the configuration interaction singles (CIS) calculations were performed using the ACES ZZ program package.33 The MP2 (19)Gauss, J. Chem. Phys. Lett. 1992, 191, 614. (20)Gauss, J. J. Chem. Phys. 1993, 99, 3629. (21)Gauss. J. Chem. Phys. Lett. 1994, 229, 198. (22)Gauss, J.; Stanton, J. F. J. Chem. Phys. 1995, 102, 251. (23)Gauss, J.; Stanton, J. F. J. Chem. Phys. 1995, 103, 3561. (24) Stanton, J. F.; Bartlett, R. J. J . Chem. Phys. 1992, 97, 2299. (25) Schafer, A.; Hom, H.; Ahlrichs, R. J. Chem. Phys. 1992,97,2571. (26) Schafer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994,100,5829. (27) Dunning, T. H., Jr. J . Chem. Phys. 1989, 90, 1007. (28) Kendall, R. A,; Dunning, T. H., Jr.; Harrison, R. J. J. Chem. Phys. 1992, 96, 6796. (29) Purvis, G. D.; Bartlett, R. J. J. Chem. Phys. 1982, 76, 1910. (30) Raghavachari, K.; Trucks, G. W.; Pople, J. A,; Head-Gordon, M. Chem. Phys. Lett. 1989, 157, 479. (31) Bartlett, R. J.; Watts, J. D.; Kucharski, S. A,; Noga, J. Chem. Phys. Lett. 1990, 165, 513. Errata Chem. Phys. Lett. 1990, 167, 609. (32) Foresman, J.; Head-Gordon, M.; Pople, J.; Frisch, M. J . Phys. Chem. 1992, 96, 135, and references therein.

ONCSb SNCO-b HOSCN HONCS" HSNCO"

111.13 110.48 109.72

176.31 177.25 177.76

100.45 100.97 100.97 123.17 129.98

177.74 177.82 178.03 171.57 173.20

" The molecule has C, symmetry. The molecule is linear. The molecule has CI symmetry. R(H0) = 97.13 pm, a(H0S) = 106.09 deg, and the two dihedral angles a(HAB,ABC) = 86.15 and a(ABC,BCD) = -167.45 deg. The moleculehas C, symmetry. R(H0) = 96.43 pm, a(H0S) = 106.58 deg, and the two dihedral angles a(HAB,ABC) = 85.98 and a(ABC,BCD) = -171.63 deg. 'The molecule has CI symmetry. R(H0) = 96.52 pm, a(H0S) = 106.86 deg, and the two dihedral angles a(HAB,ABC) = 85.87 and a(ABC,BCD) = -171.09 deg. ,The molecule has C, symmetry. R(H0) = 96.27 pm, a(H0N) = 101.69 deg. g The molecule has C, symmetry. R(HS) = 133.26 pm, a(HSN) = 94.15 deg. TZVPP geometry optimizations and also CIS calculations were performed using the TURBOMOLE program package.34

3. Results and Discussion 3.1. Geometries and Energies. The equilibrium molecular structures have been optimized at the MP2 level using the DZP, TZ2P, and TZVPP basis sets. The optimized geometries are given in Table 1. The ONCS- and SNCO- molecules are linear, and the corresponding acids have C, symmetry. The OSCNisomer has C, symmetry, and its acid form has CIsymmetry. For the anions, the equilibrium bond lengths obtained in the MP2 TZ2P calculations are 0.23-1.54 pm shorter than those of the MP2 DZP structures. The only exception is R(N0) in ONCS- which is 0.62 pm longer at the MP2 TZ2P level. The largest difference in the bond lengths of the optimized MP2 DZP and MP2 TZ2P structures (1.54 pm) was found for R(S0) in OSCN-. The largest difference in the bond angles (1.16 deg) of the DZP and TZ2P calculations was found for a(SCN) in OSCN-. By adding f-functions to the basis set the sulfur bonds become slightly shorter. Comparison of the TZ2P and TZVPP structures shows that the R(SC) bond in bent OSCN- is 1.28 pm shorter for the TZVPP geometry. For SNCO-, the R(SN) bond is 0.91 pm shorter, and for ONCS- the R(SC) bond is 0.71 pm shorter in the MP2 TZVPP calculation. The other bonds and angles are not much affected by the addition of f-functions to the basis sets. The R(S0) bond for OSCN- is about 10 ppm shorter than for HOSCN, and the R(N0) bond in ONCS- is about 15 pm shorter than for HONCS. This can (33) ACES 11,an ab initio program system, authored by J. F. Stanton, J. Gauss, J. D. Watts, W. J. Lauderdale, and R. J. Bartlett, The package also contains modified versions of the MOLECULE Gaussian integral program of J. Almlof and P. R. Taylor, the ABACUS integral derivative program written by T. Helgaker, H. J. Aa. Jensen, P. Jergensen, and P. R. Taylor, and the PROPS property integral code of P. R. Taylor. See, also: Stanton, J. F.; Gauss, J.; Watts, J. D.; Lauderdale, W. J.; Bartlett, R. J. Intern. J. Quantum Chem. Symp. 1992, 26, 879. (34) TURBOMOLE, an ab initio program system, authored by R. Ahlrichs, M. B&, H. P. Baron, M. Ehrig, F. Hase, M. Haser, H. Hom, C. Kolmel, A. Schafer, U. Schneider, P. Weis, and H. Weiss. See, also: Ahlrichs, R.; B&, M.; Haser, M.; Hom, H.; Kolmel, C. Chem. Phys. Lett. 1989, 162, 165.

W Spectra f o r Hypothiocyanite and Its Isomers

J. Am. Chem. Soc., Vol. 117, No. 46, 1995 11525

Table 2. Calculated Vibrational Frequencies (in cm-I) for Three Structures of the Hypothiocyanite Anion and for One Structure of the Hypothiocyanous Acid' molecule OSCN-" HOSCNb ONCS-' ONCS-'.d SNCOSNCO-'.d

VI

v2

v3

v4

v5

v6

183 184 201 264 93 123

367 329 331 556 495 535

464 459 622 609 534 653

595 475 1272 1152 1349 1496

942 693 2230 2618 2444 2622

2053 785

The irreducible representations of the vibrations are A', A", A', A', A', and A', respectively. v7 = 1201 cm-l, vg = 2169 cm-I, and vg = 3791 cm-I. The irreducible representations of the vibrations are n, n, u, u, and u, respectively. Vibrational frequencies calculated at HF level using 6-31GC basis set.I7 e The vibrational frequencies are obtained at the MP2 level using the DZP basis set.

be explained by the presence of some double-bond character in the R(S0) and R(N0) bonds of the anions. Since the vibrational frequencies calculated at the MP2 DZP level were all real, the obtained structures are local or global minima. The calculated vibrational frequencies are given in Table 2. The total energies of the three structures for the hypothiocyanite anion and for the corresponding hypothiocyanous acids are given in Table 3, and the relative energies with respect to the linear SNCO- isomer are given in Table 4. The energies have been calculated for the optimized MP2 geometries given in Table 1. As seen in Table 3, the lowest energy is obtained for the linear SNCO- structure. Since the SNCOisomer has C and N interchanged as compared to SCN-, one may ask whether lactoperoxidase is able to isomerize thiocyanate to the lowest isomer structure or one of the higher-lying structures is generated. At the CCSD TZ2P level the linear ONCS- isomer is 158 kJ/mol above SNCO-, and the bent OSCN- is 101 kJ/mol above SNCO-. The effect of triple excitations is estimated by performing CCSD(T) calculations using the DZP basis set. The corrections for triple excitations are -8 and -3.5 kJ/mol for ONCS- and OSCN-, respectively. The addition of f-functions to the basis sets decreases the energy differences between the SNCO- and OSCN- structures and between the SNCO- and ONCS- structures by 12.1 and 4 kJ/ mol, respectively. By adding the triple excitation contribution and the f-function contribution to the CCSD TZ2P energies, the extrapolated CCSD(T) TZVPP energies of 86 and 146 kJ/ mol relative to SNCO- are obtained for OSCN- and ONCS-, respectively. Pyykko and Runeberg" obtained also the lowest energy for the SNCO- structure. In their MP2 6-31G* calculations, the linear ONCS- was 145 kJ/mol above SNCO-. However, they did not find any stable structure for OSCN-. The proton affinity for the bent OSCN- structure calculated at the MP2 TZ2P level is 1462 kT/mol. By including f-functions

in the basis sets the obtained proton affinity calculated at the M E ! level becomes 1439 kJ/mol. The proton affinities obtained at the MP2 DZP, CCSD DZP, and CCSD(T) DZP levels are 1518, 1530, and 1528 kJ/mol, respectively, yielding an extrapolated CCSD(T) TZVPP value of 1452 kJ/mol for the proton affinity of OSCN-. The proton affinities for ONCS- (HONCS) and SNCO- (HSNCO) calculated at the MP2 TZVPP level are 1381 and 1417 kJ/mol, respectively. At the SCF level SNCOH is about 50 kJ/mol above HSNCO. The dissociation limit for SO(3Z) CN-(lZ) obtained at the MP2 TZ2P level is 258 kJ/mol above SNCO-, and the dissociation to SO('Z) CN-('Z) is further about 100 kI/mol higher up in energy. The bent OSCN- is thus about 150 and 250 kJ/mol below the dissociation limits. 3.2. The 15NChemical Shieldings. The experimental values for the I5N chemical shifts in the presence of 0.1 pM lactoperoxidase in a 0.1 M phosphate buffer solution (pH = 6.1) are -128, -170, and -179 ppm for CN-, SCN-, and OSCN-, respe~tive1y.l~An external N03- aqueous solution was used as reference. The I5N chemical shift for a pure 0.3 M N03aqueous solution has been measured to be -3.55 ppm35 with respect to neat CH3N02. Since the absolute chemical shielding for neat CH~NOZis -135.8 the experimental I5N chemical shieldings for CN-, SCN-, and OSCN- in the phosphate buffer solution are -4.3, 37.7, and 46.7 ppm, respectively. Witanowski et al.35have measured the I5N chemical shifts for 0.3 M aqueous solutions of KCN and KSCN with CH3N02 as reference. They obtained experimental values of -106.11 ppm for KCN and - 174.07 ppm for KSCN which correspond to absolute chemical shieldings of -29.7 and 38.3 ppm for KCN and KSCN, respectively. Howarth et al.37and Vaes et al.38have reported I4N and I5N chemical shifts of -166 and -159.5 ppm for SCN- with respect to a NO3- reference. The values for the I5N chemical shielding for SCN- in the aqueous solution and for SCN- in the phosphate buffer solution agree rather well. However, the 15N chemical shielding of -29.7 ppm for an aqueous solution of KCN obtained by Witanowski et al.35 is 25 ppm smaller than the I5N chemical shielding for CN- in the phosphate buffer s01ution.I~ Cyanide is known to bind to iron in lactoperoxidase, but the I5N resonance of the lactoperoxidase-cyanide complex has been found at f 4 2 3 ppm relative to N03-,39 so the I5N chemical shielding for CN- must be shifted 25 ppm by the presence of the phosphate buffer. The I5N chemical shielding for SCN- is almost unaffected by the presence of the phosphate buffer. The shift of the I5N chemical shieldings for HOSCN/OSCNintroduced by the presence of the phosphate buffer may also be a few tens of ppm.

+

+

Table 3. Energies (in au) of the Three Structures of the Hypothiocyanite Anion and of the Corresponding Structure of the Hypothiocyanous Acids as a Function of the Size of the Basis Set and of the Level of Correlation" molecule

basis set

SCF

OSCN-

DZP TZ2P TZVPP DZP TZ2P TZVPP DZP TZ2P TZVPP DZP TZ2P TZVPP TZVPP TZVPP

-564.624 840 -564.739 035 -564.760 175 -564.611 913 -564.713 466 -564.731 21 1 -564.687 587 -564.787 845 -564.804 560 -565.197 654 -565.300 247 -565.316 214 565.267 020 -565.347 509

ONCSSNCOHOSCN HONCS HSNCO

The molecular structures are optimized at the MP2 level.

MP2 -565.380 -565.616 -565.667 -565.377 -565.599 -565.647 -565.433 -565.654 -565.700 -565.957 -566.173 -566.215 -566.173 -566.240

127 238 396 509 610 667 413 187 732 854 005 600 774 297

CCSD

CCSD(T)

-565.391 914 -565.619 822

-565.41 8408

-565.383 007 -565.598 204

-565.41 1085

-565.444 464 -565.658 435

-565.469 627

-565.975 464

-566.000868

11526 J. Am. Chem. SOC., Vol. 11 7, No. 46, 1995

Sundholm

Table 4. Relative Energies (in kJ/mol)" of the Three Structures of

Table 6. Calculated Chemical Shieldings for the Hypothiocyanite

the Hypothiocyanite Anion and of One Structure of the Hypothiocyanous Acid as a Function of the Size of the Basis Set and of the Level of Correlationb

Anions and for the Hypothiocyanous Acid as a Function of the Level of Correlation and the Size of the Basis Set (in ppm)

molecule

basis set

SCF

MP2

OSCN-

DZP TZ2P TZVPP DZP TZ2P TZVPP DZP TZ2P TZVPP

164.7 128.2 116.5 198.7 195.3 192.6 -1339.2 - 1345.3 -1343.3

139.9 99.6 87.5 146.8 143.3 139.3 -1376.9 - 1362.2 - 1351.8

ONCSHOSCN

CCSD

CCSD(T)

138.0 101.4

134.5

161.4 158.1

153.7

-1394.1

-1394.8

molecule OSCN-

HOSCN ONCS-

1 au = 2625.5 kJ/mol. The molecular structures are optimized at the MP2 level. The energies of the linear SNCO- isomer are used as references. (I

SNCO-

Table 5. Comparison of Calculated and Experimental I5N Chemical Shieldings with Previously Calculated and Experimental Results (in ppm) molecule CN-

HCN

NH3

SCN-

method

a(I5N)

ref

SCF T Z 2 P MP2 T Z 2 P CCSD T Z 2 P experimentalc%d experimental',' SCF T Z 2 P MP2 T Z 2 P MP2 PZ3D2Fb CCSD PZ3D2Fb experimentalC SCF T Z 2 P MP2 T Z 2 P MP2 PZ3S2Fb CCSD PZ3D2Fb experimental' SCF T Z 2 P MP2 T Z 2 P CCSD T Z 2 P experimentaFd experimental','

-71.8 -16.4 -31.0 -4.3 -29.7 -46.4 5.9 -0.3 -16.7 -20.4 268.3 281.9 276.5 269.7 264.5 10.2 56.0 39.2 37.7 38.3

PW PW PW 14 35 PW PW 19-21 23 36 PW PW 19-21 23 40 PW PW PW 14 35

The molecular structures are optimized at the MP2 level. The experimental structures are used. Uncorrected for rovibrational motions. For rovibrational corrections see ref 21 and references therein. The chemical shift is measured in a 0.1 M phosphate buffer solution in the presence of 0.1 pM lactoperoxidase at pH = 6.1. e Measured in a 0.3 M KCN aqueous solution.

In Table 5, the calculated absolute chemical shieldings for CN-, HCN, N H 3 , and SCN- are compared to previously c a l c ~ l a t e d ~ and ~ - ~measured14~35~36~40 ~,~~ values. The basis set truncation errors are about 6 ppm which can be estimated by comparing the MP2 TZ2P and MP2 PZ3D2F values for the I5N chemical shieldings of HCN and NH3.'9-21*23 The geometrical and rovibrational corrections to the calculated chemical shieldings are of the same order of magnitude.*' The triple and higher-order excitations which are neglected in the present calculations may also contribute to the chemical shieldings. However, the agreement between the MP3, MP4, and CCSD values for the I5N chemical shieldings of HCN2lwz3and the close agreement with experiment for the I5N chemical shieldings of the molecules in Table 5 show that no large corrections due to higher order excitations can be expected for the CCSD I5N chemical shielding of OSCN-. Judging from the data in Tables 5 and 6, the uncertainty in the CCSD I5N chemical shielding for OSCN- is less than 20 ppm. The corrections due to vibrational motions may be several (35) Witanowski, M.; Stefaniak, L.; Szymanski, S.; Januszewski, H. J. M a p . Reson. 1977,28, 217. (36) Jameson, C. L.; Jameson, A. K.; Oppusunggu, D.; Wille, S.; Burrell, P. M.; Mason, J. J. Chem. Phys. 1981, 74, 81.

HOSCNIOSCN-

calculation" SCF DZPb MP2 DZPb CCSD DZPb SCF TZ2Pb MP2 TZ2Pb CCSD TZ2Pb CCSD T Z 2 P SCF DZPd MP2 DZPd SCF D Z P MP2 D Z P SCF TZ2P MP2 TZ2P SCF D Z P MP2 D Z P SCF T Z 2 P MP2 TZ2Pe experimenef

o(I3C) o(I5N)

a(I7O) u ( ~ ~ S )

59.4 67.2 70.6 53.1 51.3 56.9 55.2 88.5 83.8 123.4 157.3 118.2 149.4 78.4 98.7 63.6 80.1

447.0 483.1 453.5 468.6 516.2 481.1 463.0 350.4 369.9 221.4 247.5 218.2 241.4 273.3 288.8 269.6 281.9

-68.7 34.3 -12.0 -86.7 3.2 -39.9 -42.0 -117.1 26.9 18.5 113.9 67.9 93.6 232.8 253.2 234.7 252.2 46.7

301.8 272.2 272.2 271.2 232.5 255.8 244.0 178.4 179.3 797.3 830.6 809.5 840.6 159.1 255.4 223.0 309.3

a The molecular structures are optimized at the MP2 level. The molecule has C, symmetry. Calculated in the DZP geometry using TZ2P basis set. dThe proton shieldings are 28.8 and 29.6 ppm as obtained in the SCF DZP and MP2 DZP calculations, respectively. e The molecule is linear. fThe measured chemical shielding for hypothiocyanite and hypothiocyanous acid. See text.

ppm. The I5N chemical shieldings obtained using the TZ2P basis sets and the DZP and TZ2P structures differ by only 2 ppm. This shows that the I5N chemical shielding for OSCNis not very sensitive to nuclear displacements. However, for comparison with experiment the largest uncertainty is introduced by the presence of the buffer solution. For the molecules studied here, the I5N chemical shieldings obtained at the SCF level are too small, and the correlation corrections are overestimated at the MP2 level. See Tables 5 and 6 . An extrapolated value for the I5N chemical shieldings may therefore be obtained by subtracting one-third of the correlation correction obtained at the MP2 level from the MP2 chemical shielding. See, for example, the tables in refs 20 and 22. The extrapolated I5N chemical shieldings for ONCS- and SNCO- become 85 and 246 ppm, respectively. Remarkable is that for the I5N chemical shielding of CN-, the medium shift is 25 ppm upfield, but for SCN- there is almost no shift due to the presence of the lactoperoxidase and phosphate solution. The solvent induced shifts of the I5N resonances correlate with the hydrogen-bonding capacity of the nitrogen lone-pair electron^.^^ For hydrogen-bonding solvents, the I5N resonance is shifted up field^.^^ For the bent OSCN- structure, the difference between the CCSD TZ2P value for the I5N chemical shielding and the experimental result is 86 ppm. Assuming that the hypothiocyanite anion possesses the bent structure, the medium shift would in this case be in the same direction as for CN- (upfield), and about three times larger than for CN-. Assuming a linear ONCS- structure, the calculated I5N chemical shielding would be about 40 ppm larger than the experimental value, and the shift due to the buffer solution would for ONCS- be downfield, i.e., in the opposite direction as compared to CN-. The extrapolated value for the I5N chemical shielding for SNCO- is about 200 ppm larger than the experimental result. This discrepancy is too large to be explained by solvent and vibrational effects, and one may conclude that the SNCOstructure is not produced in the lactoperoxidase oxidation reaction. The solvent induced shifts of the I5N resonances were estimated at the SCF level using DZP basis set. The structures

W Spectra for Hypothiocyanite and Its Isomers of HSNCO, HONCS, and HOSCN surrounded by six water molecules were optimized at the SCF level. In the optimization all structural parameters were optimized. The solvent induced shifts were estimated as the difference in the chemical shieldings obtained with and without the water molecules but with unrelaxed molecular structure. The obtained solvent shifts for HSNCO, HONCS, and HOSCN were -9, -5, and +27 ppm, respectively. For HOSCN, the obtained solvent induced shift reduces the discrepancy between the calculated and measured chemical shieldings by 27 ppm which is about one-third of the discrepancy between the calculated and experimental values. The shift due to the presence of the phosphate buffer is not considered in the calculation. For HSNCO and HONCS, which have nitrogen in the middle of the molecule, the solvent induced shifts were found to be significantly smaller. The I5N chemical shieldings for HOSCN and OSCNobtained at the MP2 DZP level are quite similar. In the measurement, Modi et al.I4 could not find any difference in the 15N chemical shifts for OSCN- and HOSCN. Since the pK, value is 5.3 for the hypothiocyanous acid: both forms are present at pH = 6.1. They did not find any pH dependency of the position of the I5N resonance which shows that the "N chemical shieldings for the acid and base forms are equal. The nitrogen environment is quite similar for the hypothiocyanite anion and for the hypothiocyanous acid. The protonation of OSCN- does not significantly affect the SCN structure. This is supported by the present calculations, see Table 1. 3.3. The I3C, 1 7 0 , and 33SChemical Shieldings. The I3C, 170,and 33Schemical shieldings have been calculated for the three hypothiocyanite structures and for the bent structure of the hypothiocyanousacid at the SCF and MP2 levels using DZP and TZ2P basis sets. The chemical shieldings of the bent hypothiocyanite structure has also been performed at the CCSD level. The results are summarized in Table 6. The I3C, I7O, and 33S chemical shieldings have not been measured for hypothiocyanite. Comparison of the I3C chemical shieldings shows that at least the ONCS- structure would be distinguished from the others by measuring the I3C NMR shifts. The I3C chemical shielding for ONCS- is at the MP2 TZ2P level 149 ppm. The corresponding values for OSCN- and SNCO- are 51 and 80 ppm, respectively. The solvent shifts for the I3C shieldings are in general smaller than for the I5N ~hieldings:~and the OSCNand SNCO- structures could probably be identified by 13CNMR spectroscopy. The solvent induced shifts for I3C estimated at the SCF level are -1, +3, and -5 ppm for HSNCO, HONCS, and HOSCN, respectively. The protonation shifts for the 13C chemical shieldings calculated at the MP2 DZP level is 17 ppm, and a pH dependency of the 13C chemical shielding would probably be observed. The I3C chemical shielding for CN- (R(CN) = 118.43 pm) calculated at the M E TZ2P level is 34.2 ppm, and for the SCN(R(CS) = 166.28 pm and R(NC) = 118.40 pm) it is 68.3 ppm. At the CCSD TZ2P level they are 24.9 and 59.4 ppm, respectively. The relative I3C chemical shifts with respect to tetramethylsilane (TMS) may be obtained by comparing chemical shieldings (37)Howarth. 0. W.: Richards. R. E.: Venanzi. L. M. J. Chem. SOC. 1964,3335. (38) Vaes, J.; Chabanel, M.; Martin, M. L. J . Phys. Chem. 1978, 82, 2420. (39) Behere, D. V.; Gonzalez-Vergara, E.; Goff, H. M. Biochim. Biophys. Acta 1985, 832, 319. (40) Kukolich, S . G. J . Am. Chem. SOC. 1975, 97, 5704. (41) Martin. G.3.: Martin. M. L.: Gouesnard. J. P. I5N-NMR Soecrroscopy; Springer': i9si. (42) Duthaler, R. 0.;Roberts, J. D. J . Am. Chem. SOC. 1978, 100,4969. (43) Breitmaier, E.; Voelter, W. "C NMR Spectroscopy; Verlag Chemie: 1978. ~~

~I

J. Am. Chem. SOC., Vol. 117, No. 46, I995 11527 calculated at the same level of approximation and experimental results. The experimental value for the I3C chemical shielding of TMS is 188.1 The experimental I3C chemical shift of SCN- with TMS as reference is 134.0 ppmU yielding an experimental 13C chemical shielding of 54.1 ppm for SCN-. The TMS and SCN- I3C chemical shieldings are not corrected for vibrational motion. The I3C chemical shifts for OSCN-, HOSCN, ONCS-, and SNCO-, with respect to TMS obtained by comparing the I3C chemical shieldings calculated at the MP2 TZ2P and MP2 DZP levels, are 151.0, 134.4, 52.9, and 122.2 ppm, respectively. The I7O chemical shieldings for OSCN-, ONCS-, and SNCO- calculated at the MP2 TZ2P level are 516, 241, and 282 ppm, respectively. The OSCN- structure would be distinguished from the others by measuring the I7O NMR spectra. The 170chemical shielding for HOSCN calculated at the MP2 DZP level is 113 ppm smaller than the I7O chemical shielding for OSCN-. As seen from the two CCSD calculations of the chemical shieldings, the I7O chemical shielding for OSCN- is sensitive to changes in the geometry, and one may expect that vibrational corrections are large. However, the difference of 113 ppm is probably larger than solvent and vibrational effects. At the SCF level, the estimated solvent induced 170shifts are $2, +8, and -8 ppm for HSNCO, HONCS, and HOSCN, respectively. The I7O chemical shielding of HOSCN/OSCN- will show a pH dependency from which the individual concentrations of HOSCN and OSCN- can be deduced. The 33S chemical shieldings for OSCN-, ONCS-, and SNCO- calculated at the MP2 TZ2P level are 232, 840, and 308 ppm, respectively. Since all 33Schemical shieldings for these molecules differ, the oxidation product could probably be identified with the aid of 33SNMR spectroscopy. The 33S chemical shielding for HOSCN calculated at the Mp2 DZP level is 93 ppm smaller than the 33Schemical shielding for OSCN-, and the 33S NMR spectroscopy is probably also capable of giving a measurement of the individual concentrations of HOSCN and OSCN- in mixtures of these two substances. At the SCF level, the estimated solvent induced 33Sshifts are +31, +14, and +31 ppm for HSNCO, HONCS, and HOSCN, respectively. The 33Schemical shieldings for SCN- (R(CS) = 166.28 pm and R(NC) = 118.40 pm) calculated at the MP2 TZ2P level is 824.8 ppm, and at the CCSD TZ2P level it is 806.3 ppm. 3.4. Excitation Energies. The ultraviolet (W)spectra of HOSCN/OSCN- have been measured at pH = 8 and 3 by Tenovuo et a l . I 6 The hypothiocyanite anion and the hypothiocyanous acid have significantly different UV spectra. The hypothiocyanite spectrum shows a maximum at about 220 nm. This strong band has a maximum extinction coefficient of 3870 M-' cm-' at 220 nm, and the oscillator strength estimated from the spectrum is 0.11. For the hypothiocyanous acid, the ultraviolet absorption in the range 200-300 nm is much weaker. The ultraviolet spectrum for the hypothiocyanous acid has its maximum at about 240 nm. The extinction coefficient at 240 nm is 95 M-' cm-I, and the oscillator strength is estimated from the spectrum to be 0.14 x Since the excited states of the free anions are unstable, it is not possible to calculate reliable excitation energies for the anions. However, the UV spectrum for the hypothiocyanous acid has been measured, and a comparison of calculated and measured UV spectra for the acids supports the determination of the structure of HOSCN/OSCN-. In Table 7, the calculated excitation energies for the acid forms are given and compared (44) Kargol, J. A,; Cercely, R. W.; Burmeister, J. L. Inorg. Chem. 1979, 18, 2532.

11528 J. Am. Chem. SOC., Vol. 117,No. 46, 1995

Sundholm

Table 7. Calculated Excitation Energies (in eV) and the

4. Conclusions Three molecular structures for hypothiocyanite and the corresponding structures for the hypothiocyanous acid have been optimized at the MP2 level using different basis sets. The I3C, 15N, 170,and 33Schemical shieldings have been calculated at the SCF, MP2, and CCSD levels for two linear and one bent structures of hypothiocyanite using the gauge-including atomicorbital a p p r o a ~ h . ' ~ - ~ ~ The linear SNCO- structure is found to be the energetically most stable isomer. The bent OSCN- and linear ONCS- are calculated to be 86 and 146 kJ/mol above SNCO-, respectively. This is in agreement with previous calculations by Pyykko and RunebergI7 who in MP2 6-31G* calculations also found that SNCO- is the lowest structure of hypothiocyanite and that ONCS- is 145 kJ/mol above it. However, they did not find any stable structure for OSCN-. In the present calculations, no other low-lying structures for hypothiocyanite have been found. Comparison of calculated and measured 15Nchemical shieldings shows that the linear SNCO- structure is not generated in the lactoperoxidase oxidation reaction of SCN- by H202. The calculated 15Nchemical shielding for linear SNCO- differs by about 200 ppm from the experimental result, and for linear ONCS- and bent OSCN- the difference between the calculated and measured shieldings are 40 and -86 ppm, respectively. Medium and vibrational corrections are estimated to be several tens of ppm, and the oxidation product cannot be uniquely identified from the I5N chemical shielding data. At the MP2 level, HOSCN and OSCN- have about the same I5N chemical shieldings which also is supported by the experiment.I4 The solvent shifts are in general smaller for 13C chemical shieldings than for I5N chemical shieldings, and all three isomers would probably be identified by 13CNMR spectroscopy. The pH dependency of the I7O and 33S chemical shieldings for mixtures of HOSCN and OSCN- would provide a measurement of the individual concentrations of the two substances. The first few excitation energies have been calculated for three structures of the hypothiocyanous acid using the EOMCCSD method.24 The lowest excitation energy for HONCS is found to be 2 eV (165 nm) too low in energy. The linear ONCS- isomer is also 60 kJ/mol above bent OSCN-, and we may conclude that HONCS/ONCS- is not the oxidation product of lactoperoxidase oxidation reaction of SCN- by H202. The lowest dipole allowed excitation for HSNCO is calculated to be 4.57 eV which is in qualitative agreement with experiment. However, the chemical shift calculations on the linear SNCOstructure showed that HSNCO/SNCO- is not generated in the enzymatic oxidation reaction. Since no other low-lying structures for hypothiocyanite have been found, the only remaining structure for hypothiocyanite is the HOSCN/OSCN- isomer. For HOSCN, both the calculated UV spectrum and the I5N chemical shift are in qualitative agreement with experiment. The stable oxidationproduct obtained by oxidation of aqueous solutions of KSCN by bromine or by electrolysis2may be linear SNCO- which was found to be the energetically lowest structure.

Corresponding Wavelengths (in nm), for Three Structures of the Hypothiocyanous Acid Using the EOM-CCSD Method and DZP Basis Sets as Compared to the Experimental Resultsb molecule excitation excitation energy wavelength HOSCN 1A-2A 4.57 27 1 1A - 3A 6.49 191 HONCS 1A' - 1A" 3.05 406 1A' - 2A' 4.84 256 1A' - 2A" 6.25 198 HSNCO 1A' - 1A" 4.46 278 1A' - 2A" 5.96 207 1A' - 2A' 6.74 184 HX" experiment 5.17 240 Hypothiocyanous acid. The molecular structures are optimized at the MP2 level.

to the experimental value. The excitation energies have been calculated using the EOM-CCSD method24and DZP basis set. The EOM-CCSD method is able to provide a good description of excited states dominated by single e~citations.4~The approximate excitation level (AEL)45is about 1.1 for all excited states considered in this work. For HONCS, the excitation energy for the first dipole transition is calculated to be 3.05 eV (406 nm) which is 2 eV smaller than the maximum of the experimental spectrum. The shift of the energy levels due to the solvent is much smaller than the energy difference of 2 eV between the experimental and calculated excitation energies, and we may conclude that the HONCS/ONCS- is not the generated product of the lactoperoxidase oxidation. The first dipole transition for HSNCO is obtained at 4.46 eV (278 nm). This excitation energy is only 0.7 eV smaller than the experimental value, but as we saw previously, the I5N chemical shielding for the SNCO- anion is about 200 ppm larger than the experimental result, and we concluded that linear SNCO- is not generated in the peroxidase reaction. However, the stable oxidation product obtained by the oxidation of aqueous solutions of KSCN by bromine or generated by electrolysis2 may be linear SNCO- which was found to be the energetically lowest structure. The first excitation energy for HOSCN of 4.57 eV (271 nm) agrees qualitatively with the experimental value of 5.17 eV (240 nm), which indicate that HOSCN/OSCNstructure is produced in the lactoperoxidase oxidation reaction. For the acids all studied transitions are weak, and the oscillator strengths do not provide any additional information. The effect of diffuse functions on the excitation energies was estimated at the CIS level by adding diffuse functions to the DZP basis set. For three acid structures, the excitation energies for the first excited state of each symmetry was almost unaffected (0.01-0.09 eV) by the addition of diffuse functions to the basis set. The excitation energies of the second excited states of each symmetry decreased by about 0.3 eV due to the addition of diffuse basis functions. The solvent effects on the excitation energies are somewhat more important for the first excitation energies than addition of diffuse functions. The solvent shifts of the excitation energies were estimated as follows: The acids were surrounded by six water molecules, and all structural parameters were optimized at the SCF level. The solvent shifts were obtained as the difference in the CIS excitation energies with water and without water but with unrelaxed molecular structure. The solvent effects were found to stabilize the excited states by 0.1-0.3 eV. (45) Bartlett, R. J.; Stanton, J. F. In Reviews ofCompututional Chemistry; Lipkowitz, K. B., Boyd, D. B., Eds.; VCH Publishers: New York, 1990; Vol. 5 , p 65.

Acknowledgment. The research reported here has been supported by grants from The Academy of Finland and the Alexander von Humboldt Stiftung. This work has also been supported by the Fonds der Chemischen Industrie. The author thanks Pekka Pyykko and Nino Runeberg for pointing out that the 15N chemical shifts and the UV spectrum have been measured for hypothiocyanite. The author thanks Jiirgen Gauss for performing the CCSD chemical shift calculations and for fruitful discussions.

JA9504449